fredag den 14. februar 2014

A little extension for diagnosing tax code combination errors

In Dynamics AX 2012 if General Ledger module is setup to validate Check sales tax by entering either "Warning" or "Error" in the "Check sales tax groups" field, then Dynamics AX will throw an error (stop execution) or issue a warning in the infolog if it encounters a tax setup error when doing calculation of tax.



The error looks like this:


As you might know the tax calculation is based on getting the intersecting tax codes connected to the item tax group and the tax group of a given source document line.

In this case I encountered we wanted to diagnose the origins of the error, because the error occurred when changing a customer group, which we thought was a bit odd.

Digging a bit deeper, the error was not related to changing the customer group, but instead related to the following recalculation of credit limit on the customer that Dynamics AX wanted to perform.

I traced the error to the class Tax and the method insertInterSection.

Seing that the Tax class is a super class the reason for the not very informative error message became clear. The insertIntersection method is inherited by the sub classes in the Tax class hierachy and therefor contains logic that is not module specific, so there are no references to the origin-lines (e.g. sales lines og purch lines) that the tax group set up comes from.

To get info about the origins of the error I did the following:

I added a new method to be used when an error is to be reported:

private str invalidTaxCodeCombErrortxt(ErrorTxt _errorTxt)
{
DictTable dictTable;
Common dCommon;
ErrorTxt foundInErrorTxtPostfix;

dictTable = new dictTable(this.sourceTableId());
dCommon = dictTable.makeRecord();
select firstOnly dCommon where dCommon.recid == this.sourceRecId();
switch (dCommon.TableId)
{
case tableNum(SalesLine):
foundInErrorTxtPostfix = strFmt(" found in : %1 linje %2",dCommon.(fieldNum(SalesLine,SalesId)),dCommon.(fieldNum(SalesLine,LineNum)));
break;
case tableNum(PurchLine):
foundInErrorTxtPostfix = strFmt(" found in : %1 linje %2",dCommon.(fieldNum(PurchLine,PurchId)),dCommon.(fieldNum(PurchLine,LineNumber)));
break;
default:
foundInErrorTxtPostfix = '';
}
return _errorTxt + ' ' + foundInErrorTxtPostfix;
}

The method attempts to evaluate a generic tablebuffer to determine what is the source of the tax codes being evaluated, and which cause an invalid combination.
It does so by instantiating a DictTable object using the table id found in the tax class (this.sourceTableId()). It then does a call to the kernel method makeRecord() on the same object to get a tablebuffer.
The we do a select firstonly on this tablebuffer using the recid in (this.sourceTableId()) to get the record that is the origin of the tax groups.

The it evaluates the table id of the found record and constructs and error message text, that can be post fixed to the standard error message.

To notice in the

dCommon.(fieldNum(SalesLine,SalesId))

construct which actually let's you access the fields of a common table buffer.

The InsertIntersection method is changed so that the switch case construct in line 175 looks like this:

switch (this.taxParameters().CheckIntersection)
{
    case CheckTaxGroups::Warning :
        //warning(strFmt("@SYS75213", _taxGroup, _taxItemGroup));
        warning(this.invalidTaxCodeCombErrortxt(strFmt("@SYS75213", _taxGroup, _taxItemGroup)));

        break;
    case CheckTaxGroups::Error :        //warning(strFmt("@SYS75213", _taxGroup, _taxItemGroup));
        warning(this.invalidTaxCodeCombErrortxt(strFmt("@SYS75213", _taxGroup, _taxItemGroup)));
        break;
effectively adding more info to the error message.



This way I was able to pinpoint the source of the problem: An open sales order line on the customer we edited was missing an item tax group.